Nominal Group Technique

A way of running group decisions in which people generate their opinions independently and asynchronously before any meeting, so that crosstalk cannot contaminate the inputs. A nominal group is a group that, at the moment of forming opinions, works alone. Annie Duke calls this the single easiest high-impact change a company can make, and the one she most often gets teams to actually adopt.

What goes wrong in an ordinary meeting

People assume a meeting is for three things — discover, discuss, decide — and cram all three into the room. Discovery in a group is the damaging part. When everyone calls out estimates together, the loudest voice and the most-confident voice gain outsized influence, which is fine only when they are also right. Opinions converge before they have been independently formed.

The fix: split the three

Discover, independently and beforehand. Email each participant the question and collect answers privately — ‘don’t reply all, send it back to me’. The format varies: an open brainstorm, a forced rank of options with a three-to-five-sentence rationale, or a forecast with point estimate and bounds. For repeated decisions, collect through a form that hides others’ responses (a spreadsheet only the organiser sees).

Discuss, in the room. Circulate the collected opinions, then meet only to discuss them — concentrating on where people disagree rather than double-clicking on the roughly 80% that already agrees.

Decide, outside the room. A single decision-maker (Duke’s preference) or a private vote. Even a six-person partnership votes independently and need not reach agreement.

Facilitate by reflecting, not agreeing

The facilitator never says ‘I agree’. They reflect each contribution back — ‘what I heard you say is…’ — until the speaker confirms or corrects it. People then feel genuinely heard, become endowed to the outcome (they own it), and accept not getting everything they wanted because they have seen the true spread of opinion.

Drop the word ‘alignment’

Ten different people will not leave a room agreeing, and expecting them to is both unreal and quietly coercive. The goal of speaking is to convey a view, not to convince; as soon as the goal becomes agreement, interrupting and ‘you’re wrong’ creep in and people stake out tribal ground. Independence preserves real signal: a junior ‘four weeks’ estimate survives next to a senior ‘three months’ instead of conforming to it — and each party gets to learn from the other.

You can do it on the fly

Independence does not require advance planning. When a live discussion starts converging — ‘yeah, but that’ll take…’ — Duke stops the room: ‘everybody take out a piece of paper.’ The same independent elicitation, mid-meeting.

Source

See Annie Duke on Better Decisions, Kill Criteria, and When to Quit. The First Round Capital investment rubric is the worked institutional case. Related: Kill Criteria, Thinking in Bets, OKRs.